{"id":1074,"date":"2026-06-28T05:30:36","date_gmt":"2026-06-27T20:30:36","guid":{"rendered":"https:\/\/itexplore.org\/jp\/columns\/ai-model-evolution-chip-development-distillation-threats\/"},"modified":"2026-06-28T05:30:36","modified_gmt":"2026-06-27T20:30:36","slug":"ai-model-evolution-chip-development-distillation-threats","status":"publish","type":"post","link":"https:\/\/itexplore.org\/jp\/columns\/ai-model-evolution-chip-development-distillation-threats\/","title":{"rendered":"AI\u30e2\u30c7\u30eb\u9032\u5316\u3001\u30c1\u30c3\u30d7\u958b\u767a\u7af6\u4e89\u3001\u305d\u3057\u3066\u84b8\u7559\u653b\u6483\u306e\u8105\u5a01"},"content":{"rendered":"<p>\u672c\u65e5\u306e\u6ce8\u76eeAI\u30fb\u30c6\u30c3\u30af\u30cb\u30e5\u30fc\u30b9\u3092\u3001\u5c02\u9580\u7684\u306a\u5206\u6790\u3068\u5171\u306b\u304a\u5c4a\u3051\u3057\u307e\u3059\u3002<\/p>\n<div class=\"wp-block-vk-blocks-alert vk_alert alert alert-warning has-alert-icon\">\n<div class=\"vk_alert_icon\">\n<div class=\"vk_alert_icon_icon\"><i class=\"fa-solid fa-triangle-exclamation\" aria-hidden=\"true\"><\/i><\/div>\n<div class=\"vk_alert_icon_text\"><span>Warning<\/span><\/div>\n<\/div>\n<div class=\"vk_alert_content\">\n<p>\u3053\u306e\u8a18\u4e8b\u306fAI\u306b\u3088\u3063\u3066\u81ea\u52d5\u751f\u6210\u30fb\u5206\u6790\u3055\u308c\u305f\u3082\u306e\u3067\u3059\u3002AI\u306e\u6027\u8cea\u4e0a\u3001\u4e8b\u5b9f\u8aa4\u8a8d\u304c\u542b\u307e\u308c\u308b\u53ef\u80fd\u6027\u304c\u3042\u308b\u305f\u3081\u3001\u91cd\u8981\u306a\u5224\u65ad\u3092\u4e0b\u3059\u969b\u306f\u5fc5\u305a\u30ea\u30f3\u30af\u5148\u306e\u4e00\u6b21\u30bd\u30fc\u30b9\u3092\u3054\u78ba\u8a8d\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/div>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">GPT-5.6 Sol\uff1a\u6b21\u4e16\u4ee3\u30e2\u30c7\u30eb\u306e\u30d7\u30ec\u30d3\u30e5\u30fc<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Previewing GPT-5.6 Sol: a next-generation model<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>OpenAI<\/strong>\u306f\u3001\u6b21\u4e16\u4ee3\u30e2\u30c7\u30eb\u3067\u3042\u308b<strong>GPT-5.6<\/strong>\u30b7\u30ea\u30fc\u30ba\uff08<strong>Sol<\/strong>\u3001<strong>Terra<\/strong>\u3001<strong>Luna<\/strong>\uff09\u306e\u9650\u5b9a\u30d7\u30ec\u30d3\u30e5\u30fc\u3092\u958b\u59cb\u3057\u307e\u3057\u305f\u3002\u3053\u308c\u3089\u306e\u30e2\u30c7\u30eb\u306f\u3001\u7279\u306b\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u3001\u751f\u7269\u5b66\u3001\u30b5\u30a4\u30d0\u30fc\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u306b\u304a\u3051\u308b\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u80fd\u529b\u304c\u5927\u5e45\u306b\u5411\u4e0a\u3057\u3066\u3044\u307e\u3059\u3002<strong>GPT-5.6 Sol<\/strong>\u306f\u3001\u6700\u3082\u5f37\u529b\u306a\u30e2\u30c7\u30eb\u3068\u3057\u3066\u4f4d\u7f6e\u3065\u3051\u3089\u308c\u3001<strong>Terminal-Bench 2.1<\/strong>\u3084<strong>GeneBench v1<\/strong>\u3001<strong>ExploitBench<\/strong>\u306a\u3069\u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3067\u6700\u5148\u7aef\u306e\u6027\u80fd\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u65b0\u6a5f\u80fd\u3068\u3057\u3066\u3001<strong>Sol<\/strong>\u306b\u306f\u3088\u308a\u6df1\u304f\u63a8\u8ad6\u3059\u308b\u305f\u3081\u306e\u300c\u6700\u5927\u63a8\u8ad6\u52aa\u529b\u300d\u3068\u3001\u30b5\u30d6\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3092\u6d3b\u7528\u3057\u3066\u8907\u96d1\u306a\u4f5c\u696d\u3092\u52a0\u901f\u3059\u308b\u300c\u30a6\u30eb\u30c8\u30e9\u30e2\u30fc\u30c9\u300d\u304c\u5c0e\u5165\u3055\u308c\u307e\u3057\u305f\u3002 \u307e\u305f\u3001<strong>OpenAI<\/strong>\u306f\u3001\u30e2\u30c7\u30eb\u306e\u80fd\u529b\u5411\u4e0a\u306b\u5408\u308f\u305b\u3066\u3001\u3088\u308a\u5805\u7262\u306a\u591a\u5c64\u7684\u306a\u30bb\u30fc\u30d5\u30ac\u30fc\u30c9\u30b9\u30bf\u30c3\u30af\u3092\u5b9f\u88c5\u3057\u3001\u4e0d\u6b63\u5229\u7528\u3092\u6291\u5236\u3057\u3064\u3064\u3001\u6b63\u5f53\u306a\u9632\u5fa1\u7684\u4f5c\u696d\u3078\u306e\u30a2\u30af\u30bb\u30b9\u3092\u7dad\u6301\u3059\u308b\u3053\u3068\u3092\u76ee\u6307\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u30bb\u30fc\u30d5\u30ac\u30fc\u30c9\u306b\u306f\u3001\u30e2\u30c7\u30eb\u30ec\u30d9\u30eb\u306e\u62d2\u5426\u3001\u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u306e\u60aa\u7528\u5206\u985e\u5668\u3001\u30a2\u30ab\u30a6\u30f3\u30c8\u30ec\u30d9\u30eb\u306e\u30ec\u30d3\u30e5\u30fc\u3001\u304a\u3088\u3073\u5dee\u5225\u5316\u3055\u308c\u305f\u30a2\u30af\u30bb\u30b9\u304c\u542b\u307e\u308c\u307e\u3059\u3002 <strong>OpenAI<\/strong>\u306f\u300170\u4e07\u6642\u9593\u4ee5\u4e0a\u306e<strong>A100<\/strong>\u76f8\u5f53\u306eGPU\u6642\u9593\u3092\u81ea\u52d5\u30ec\u30c3\u30c9\u30c1\u30fc\u30df\u30f3\u30b0\u306b\u8cbb\u3084\u3057\u3001\u666e\u904d\u7684\u306a\u30b8\u30a7\u30a4\u30eb\u30d6\u30ec\u30a4\u30af\u653b\u6483\u3092\u898b\u3064\u3051\u3066\u5bfe\u7b56\u3092\u5f37\u5316\u3057\u307e\u3057\u305f\u3002 <strong>GPT-5.6<\/strong>\u30b7\u30ea\u30fc\u30ba\u306f\u3001\u4eca\u5f8c\u6570\u9031\u9593\u3067\u5e83\u304f\u5229\u7528\u53ef\u80fd\u306b\u306a\u308b\u4e88\u5b9a\u3067\u3001<strong>Sol<\/strong>\u306f\u5165\u529b\u30c8\u30fc\u30af\u30f3\u3042\u305f\u308a5\u30c9\u30eb\u3001\u51fa\u529b\u30c8\u30fc\u30af\u30f3\u3042\u305f\u308a30\u30c9\u30eb\u3067\u63d0\u4f9b\u3055\u308c\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/openai.com\/index\/previewing-gpt-5-6-sol\/\" target=\"_blank\" rel=\"noopener\">OpenAI \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> OpenAI's GPT-5.6 series (Sol, Terra, Luna) introduces advanced agentic capabilities in coding, biology, and cybersecurity, backed by robust, layered safety measures and a new pricing structure, signaling a significant leap in frontier AI models.<\/li>\n<li><strong>\u8457\u8005:<\/strong> OpenAI Editorial Staff<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>OpenAI<\/strong> has initiated a limited preview of its next-generation <strong>GPT-5.6<\/strong> series, comprising <strong>Sol<\/strong>, <strong>Terra<\/strong>, and <strong>Luna<\/strong> models. These models demonstrate significantly enhanced agentic capabilities, particularly across coding, biology, and cybersecurity domains. <strong>GPT-5.6 Sol<\/strong>, positioned as the flagship model, achieves state-of-the-art performance on benchmarks like <strong>Terminal-Bench 2.1<\/strong>, <strong>GeneBench v1<\/strong>, and <strong>ExploitBench<\/strong>.<\/p>\n<p>New features include a \"max reasoning effort\" for <strong>Sol<\/strong> to enable deeper reasoning and an \"ultra mode\" that leverages subagents to accelerate complex tasks. Concurrently, <strong>OpenAI<\/strong> has implemented its most robust, layered safeguard stack to date, designed to mitigate misuse while preserving legitimate access for defensive work.<\/p>\n<p>These safeguards encompass model-level refusals, real-time misuse classifiers, account-level reviews, and differentiated access. <strong>OpenAI<\/strong> dedicated over 700,000 <strong>A100<\/strong>-equivalent GPU hours to automated red-teaming to identify and address universal jailbreak attacks. The <strong>GPT-5.6<\/strong> series is slated for broader availability in the coming weeks, with <strong>Sol<\/strong> priced at $5 per 1M input tokens and $30 per 1M output tokens.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">Claude\u306b\u5bfe\u3059\u308b\u53f2\u4e0a\u6700\u5927\u306e\u84b8\u7559\u653b\u6483\u304cAlibaba\u306b\u6ce2\u53ca\u3057\u3001\u7c73\u56fd\u3067\u8b66\u9418\u3092\u9cf4\u3089\u3059<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> El mayor ataque de destilaci\u00f3n contra Claude salpica a Alibaba y enciende las alarmas en Estados Unidos<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>Anthropic<\/strong>\u306f\u3001\u4e2d\u56fd\u306e\u5de8\u5927\u4f01\u696d<strong>Alibaba<\/strong>\u3068\u305d\u306e<strong>Qwen<\/strong>\u30e9\u30dc\u306b\u95a2\u9023\u3059\u308b\u30aa\u30da\u30ec\u30fc\u30bf\u30fc\u304c\u3001\u540c\u793e\u306e<strong>Claude AI<\/strong>\u30e2\u30c7\u30eb\u306b\u5bfe\u3057\u3066\u53f2\u4e0a\u6700\u5927\u306e\u300c\u84b8\u7559\u653b\u6483\u300d\u3092\u884c\u3063\u305f\u3068\u544a\u767a\u3057\u307e\u3057\u305f\u3002 2026\u5e744\u6708\u304b\u30896\u6708\u306e\u9593\u306b\u3001\u7d0425,000\u306e\u4e0d\u6b63\u30a2\u30ab\u30a6\u30f3\u30c8\u3092\u4f7f\u7528\u3057\u30662,880\u4e07\u56de\u3082\u306e\u30a4\u30f3\u30bf\u30e9\u30af\u30b7\u30e7\u30f3\u304c\u884c\u308f\u308c\u3001<strong>Claude<\/strong>\u306e\u9ad8\u5ea6\u306a\u63a8\u8ad6\u3001\u30d7\u30ed\u30b0\u30e9\u30df\u30f3\u30b0\u3001\u304a\u3088\u3073\u9577\u671f\u9593\u306e\u30bf\u30b9\u30af\u89e3\u6c7a\u80fd\u529b\u304c\u62bd\u51fa\u3055\u308c\u305f\u3068\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>AI\u306b\u304a\u3051\u308b\u300c\u84b8\u7559\u300d\u81ea\u4f53\u306f\u3001\u3088\u308a\u5c0f\u578b\u3067\u5b89\u4fa1\u306a\u30e2\u30c7\u30eb\u3092\u5f37\u529b\u306a\u30e2\u30c7\u30eb\u306e\u5fdc\u7b54\u3067\u8a13\u7df4\u3059\u308b\u5408\u6cd5\u7684\u306a\u6280\u8853\u3067\u3059\u304c\u3001<strong>Anthropic<\/strong>\u306f\u3001\u3053\u306e\u30b1\u30fc\u30b9\u3067\u306f\u5927\u898f\u6a21\u304b\u3064\u7121\u8a31\u53ef\u3067\u884c\u308f\u308c\u3001\u5229\u7528\u898f\u7d04\u306b\u9055\u53cd\u3057\u3066\u3044\u308b\u3068\u4e3b\u5f35\u3057\u3066\u3044\u307e\u3059\u3002 \u3053\u306e\u4e8b\u4ef6\u306f\u3001AI\u5206\u91ce\u306b\u304a\u3051\u308b\u6280\u8853\u7684\u3001\u7d4c\u6e08\u7684\u3001\u5730\u653f\u5b66\u7684\u306a\u7af6\u4e89\u306e\u6fc0\u5316\u3092\u6d6e\u304d\u5f6b\u308a\u306b\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p><strong>Anthropic<\/strong>\u306f\u3001\u540c\u69d8\u306e\u653b\u6483\u304c\u4ee5\u524d\u306b\u3082\u4e2d\u56fd\u306e\u4ed6\u306e\u30e9\u30dc\uff08<strong>DeepSeek<\/strong>\u3001<strong>Moonshot AI<\/strong>\u3001<strong>MiniMax<\/strong>\u306a\u3069\uff09\u306b\u3088\u3063\u3066\u884c\u308f\u308c\u305f\u3068\u6307\u6458\u3057\u3066\u304a\u308a\u3001\u4eca\u56de\u306e<strong>Alibaba<\/strong>\u306b\u5bfe\u3059\u308b\u544a\u767a\u306f\u3055\u3089\u306b\u5927\u898f\u6a21\u3067\u3059\u3002 \u7c73\u56fd\u653f\u5e9c\u306f\u3001\u4e2d\u56fd\u304c\u7c73\u56fd\u30e2\u30c7\u30eb\u306e\u80fd\u529b\u3092\u30b3\u30d4\u30fc\u3059\u308b\u3053\u3068\u3067AI\u30ea\u30fc\u30c0\u30fc\u30b7\u30c3\u30d7\u304c\u640d\u306a\u308f\u308c\u308b\u3053\u3068\u3092\u61f8\u5ff5\u3057\u3066\u304a\u308a\u3001<strong>Anthropic<\/strong>\u306f\u8b70\u4f1a\u306b\u5bfe\u3057\u3001\u72ec\u5360\u7981\u6b62\u6cd5\u306b\u62b5\u89e6\u305b\u305a\u306b\u60c5\u5831\u5171\u6709\u3092\u4fc3\u9032\u3057\u3001\u9ad8\u5ea6\u306a\u30c1\u30c3\u30d7\u306e\u8f38\u51fa\u7ba1\u7406\u3092\u5f37\u5316\u3057\u3001\u7121\u8a31\u53ef\u306e\u84b8\u7559\u3092\u884c\u3046\u30e9\u30dc\u306b\u5236\u88c1\u3092\u8ab2\u3059\u3088\u3046\u6c42\u3081\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/es.gizmodo.com\/el-mayor-ataque-de-destilacion-contra-claude-salpica-a-alibaba-y-enciende-las-alarmas-en-estados-unidos-2000243378\" target=\"_blank\" rel=\"noopener\">Gizmodo en Espa\u00f1ol \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> Anthropic's accusation against Alibaba for a large-scale \"distillation attack\" on its Claude AI highlights the intense geopolitical and economic competition in AI, raising alarms in the US about intellectual property theft and national security.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Thomas Handley<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>Anthropic<\/strong> has accused operators linked to Chinese giant <strong>Alibaba<\/strong> and its <strong>Qwen<\/strong> lab of orchestrating the largest known \"distillation attack\" against its <strong>Claude AI<\/strong> model. Between April and June 2026, approximately 25,000 fraudulent accounts allegedly generated 28.8 million interactions to extract advanced capabilities from <strong>Claude<\/strong>, including agentic reasoning, programming, and long-task resolution.<\/p>\n<p>While \"distillation\" in AI is a legitimate technique to train smaller, cheaper models using responses from more powerful ones, <strong>Anthropic<\/strong> asserts that this instance was conducted on a massive, unauthorized scale, violating terms of use. This incident underscores the escalating technological, economic, and geopolitical competition within the AI sector.<\/p>\n<p><strong>Anthropic<\/strong> previously identified similar campaigns by other Chinese labs (e.g., <strong>DeepSeek<\/strong>, <strong>Moonshot AI<\/strong>, <strong>MiniMax<\/strong>) in February, making the accusation against <strong>Alibaba<\/strong> even larger in scale. The U.S. government is concerned that China copying U.S. model capabilities could erode AI leadership, prompting <strong>Anthropic<\/strong> to urge Congress to facilitate information sharing without antitrust conflicts, strengthen export controls on advanced chips, and sanction labs engaging in unauthorized distillation.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">OpenAI\u304b\u3089SpaceX\u307e\u3067\u3001\u306a\u305c\u5404\u793e\u304c\u81ea\u793e\u30c1\u30c3\u30d7\u3092\u958b\u767a\u3057\u3001Nvidia\u3078\u306e\u5727\u529b\u3092\u5f37\u3081\u3066\u3044\u308b\u306e\u304b<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Why everyone from OpenAI to SpaceX is building their own chips (and turning up the heat on Nvidia) | TechCrunch<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p>\u3053\u306eTechCrunch\u306e\u8a18\u4e8b\u306f\u3001<strong>OpenAI<\/strong>\u3084<strong>SpaceX<\/strong>\u3068\u3044\u3063\u305f\u4e3b\u8981\u306a\u30c6\u30af\u30ce\u30ed\u30b8\u30fc\u4f01\u696d\u304c\u3001<strong>Nvidia<\/strong>\u306e\u3088\u3046\u306a\u65e2\u5b58\u306e\u30b5\u30d7\u30e9\u30a4\u30e4\u30fc\u3078\u306e\u4f9d\u5b58\u3092\u6e1b\u3089\u3059\u305f\u3081\u306b\u3001\u306a\u305c\u72ec\u81ea\u306eAI\u30c1\u30c3\u30d7\u3092\u958b\u767a\u3057\u3066\u3044\u308b\u306e\u304b\u3068\u3044\u3046\u30c8\u30ec\u30f3\u30c9\u3092\u6398\u308a\u4e0b\u3052\u3066\u3044\u307e\u3059\u3002 \u8a18\u4e8b\u306e\u5177\u4f53\u7684\u306a\u5185\u5bb9\u306f\u30a2\u30af\u30bb\u30b9\u3067\u304d\u307e\u305b\u3093\u3067\u3057\u305f\u304c\u3001\u3053\u306e\u52d5\u304d\u306f\u3001AI\u30e2\u30c7\u30eb\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3068\u63a8\u8ad6\u306b\u304a\u3051\u308b\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u306e\u6700\u9069\u5316\u3001\u30b3\u30b9\u30c8\u306e\u524a\u6e1b\u3001\u304a\u3088\u3073\u30b5\u30d7\u30e9\u30a4\u30c1\u30a7\u30fc\u30f3\u306e\u3088\u308a\u53b3\u5bc6\u306a\u5236\u5fa1\u3092\u76ee\u7684\u3068\u3057\u3066\u3044\u308b\u3068\u8003\u3048\u3089\u308c\u307e\u3059\u3002<\/p>\n<p>\u591a\u304f\u306e\u4f01\u696d\u304c\u3001\u6c4e\u7528GPU\u3067\u306f\u6e80\u305f\u305b\u306a\u3044\u7279\u5b9a\u306eAI\u30ef\u30fc\u30af\u30ed\u30fc\u30c9\u306e\u8981\u4ef6\u306b\u5bfe\u5fdc\u3059\u308b\u305f\u3081\u3001\u30ab\u30b9\u30bf\u30e0\u30b7\u30ea\u30b3\u30f3\u306e\u8a2d\u8a08\u306b\u6295\u8cc7\u3057\u3066\u3044\u307e\u3059\u3002 \u3053\u308c\u306f\u3001AI\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u5e02\u5834\u306b\u304a\u3051\u308b\u7af6\u4e89\u3092\u6fc0\u5316\u3055\u305b\u3001<strong>Nvidia<\/strong>\u306e\u512a\u4f4d\u6027\u306b\u6311\u6226\u3059\u308b\u53ef\u80fd\u6027\u306e\u3042\u308b\u91cd\u8981\u306a\u30b7\u30d5\u30c8\u3092\u793a\u5506\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/techcrunch.com\/video\/why-everyone-from-openai-to-spacex-is-building-their-own-chips-and-turning-up-the-heat-on-nvidia\/\" target=\"_blank\" rel=\"noopener\">TechCrunch \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> Major tech companies are increasingly developing custom AI chips to reduce reliance on Nvidia, optimize performance, and control costs, indicating a strategic shift in the AI hardware market.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Theresa Loconsolo<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p>This TechCrunch article explores the growing trend of major technology companies, including <strong>OpenAI<\/strong> and <strong>SpaceX<\/strong>, developing their own custom AI chips to reduce reliance on established suppliers like <strong>Nvidia<\/strong>. While the full content of the article was inaccessible, this strategic shift is understood to be driven by the desire to optimize performance for specific AI workloads, lower operational costs, and gain tighter control over their supply chains for AI model training and inference.<\/p>\n<p>Many companies are investing in designing custom silicon to meet the unique demands of their AI applications, which generic GPUs may not fully address. This trend signifies a significant shift in the AI hardware landscape, intensifying competition and potentially challenging <strong>Nvidia<\/strong>'s dominant position.<\/p>\n<\/blockquote>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>OpenAI\u306eGPT-5.6\u30b7\u30ea\u30fc\u30ba\u306e\u9032\u5316\u3001Anthropic\u306eClaude\u306b\u5bfe\u3059\u308b\u84b8\u7559\u653b\u6483\u3001\u305d\u3057\u3066AI\u4f01\u696d\u306b\u3088\u308b\u81ea\u793e\u30c1\u30c3\u30d7\u958b\u767a\u306e\u52d5\u5411\u3092\u6df1\u6398\u308a\u3057\u307e\u3059\u3002<\/p>\n","protected":false},"author":1,"featured_media":856,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"sns_share_botton_hide":"","vkExUnit_sns_title":"","_vk_print_noindex":"","sitemap_hide":"","_veu_custom_css":"","veu_display_promotion_alert":"","vkexunit_cta_each_option":"","_lightning_design_setting":[],"footnotes":""},"categories":[3],"tags":[8,475,58,201,472,16,44,57,56,473,26,15,474],"class_list":["post-1074","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-columns","tag-ai","tag-alibaba","tag-anthropic","tag-claude","tag-gpt-5-6-sol","tag-llm","tag-nvidia","tag-openai","tag-56","tag-473","tag-26","tag-474"],"veu_head_title_object":{"title":"","add_site_title":""},"_links":{"self":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/1074","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/comments?post=1074"}],"version-history":[{"count":0,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/1074\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media\/856"}],"wp:attachment":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media?parent=1074"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/categories?post=1074"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/tags?post=1074"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}